Hematopoietic stem cells are an attractive target for gene transfer but, whether in the bone marrow or mobilized into the peripheral blood, they are few in absolute numbers of usually very few relative to other more mature cells. Therefore, selection of transfected cells is essential if the e are to reconstitute hematopoiesis. The choice of a drug-resistance gene for selection has two advantages. (I) The myelo-toxicity of chemotherapy using that particular drug will be reduced. (ii) If any gene has been inserted in the vector (incis to the selectable gene), the same drug can be used to select in vivo for that gene. This project is based on the extensive experience in the project leader's laboratory with the enzyme dihydrofolate reductase (DHFR), which can confer methotrexate (MTX) resistance to practically any cell through increased expressed or through specific structural changes. Novel mutants of human DHFR recently generated by genetic engineering have been biochemically characterized, and those with the most desirable kinetic properties will be transduced through retroviruses into NIH 3T3 fibroblasts. The N2-based DC and MFG retroviral vector backbones will be compared, each one encoding the most promising DHFR mutants, and their efficiency in conferring resistance to MTX in primary hematopoietic cells will be quantitatively assessed in in vivo murine bone marrow chimeras. Expanding on recent results demonstrating increase gene transfer in human CD34+ peripheral blood stem cells when using GaLV envelope-coated retroviral particles, and substituting the Flk-2 ligand for the kit ligand in the pre-infection phase, MTX resistance of human hematopoietic cells will be characterized in vivo and in vivo in NOD/Lt-scid/scid xenochimeras. In order to purify transduced cells using immunomagnetic technology, and in order to characterize them with respect to the proportion of progenitors and of LTC-IC, dicistronic vectors that encode as cell surface marker an inactive mutant of the human nerve growth factor receptor have been recently generated and characterized. Finally, the retroviral construct yielding the best results in terms of sustained expression will be tested in patients with breast cancer, to determine efficiency of transfer and persistence of expression in humans, and the ability of mutant DHFR to act as a selectable marker for a second gene (thymidine kinase). These studies might have more general applicability because on one hand the DHFR cassette could be replaced with on conferring resistance to another relevant drug; on the other hand, MTX-resistance could be used to select for any other gene.